[pmwiki-users] Improve export-to-html recipe

[Johan Bengtsson]

Ok, I see now that $EnableRobotCloakActions was missing the dollar sign in our config.php resulting in a php parse error!

But still  --reject-regex '(.*)\?(.*)'  is useful with wget.

-Johan 

Med vänlig hälsning / Kind Regards
CHALMERS
Johan Bengtsson
Supporttekniker IT | Support Technician IT
Chalmers verksamhetsstöd | Chalmers Operations Support
IT-avdelningen | IT Office | IT-support | IT Support
+46(0)31 772 6734
Chalmers tekniska högskola | Chalmers University of Technology
Chalmers teknikpark / Sven Hultins Gata 9B
SE-412 96
Göteborg | Gothenburg
Sverige | Sweden
www.chalmers.se


________________________________________
Från: Petko Yotov <5ko at 5ko.fr>
Skickat: den 13 november 2025 15:18
Till: Johan Bengtsson
Kopia: pmwiki-users
Ämne: Re: [pmwiki-users] Improve export-to-html recipe

For known bot user agents, $EnableRobotCloakActions and $RobotActions
would cause links to be rewritten to not show ?action URLs in the HTML
source, and to return an error 403 Forbidden, not 500, if the agent
tries to access a forbidden action.

Moreover, wget is not among the default known bot user agents, so it
wouldn't activate the rewriting or blocking.

An internal server error 500 often appears when there is a problem with
one of the scripts, either an error or a PHP incompatibility in local
configuration, in a recipe, or in the core.

It may also be unrelated to PmWiki, say a web application firewall
configuration.

Are there any fatal error details in the PHP error log?

Petko

On 13/11/2025 14:38, Johan Bengtsson wrote:
> Hi,
>
> I tried https://www.pmwiki.org/wiki/Cookbook/ExportHTML but found that
> it didn't work so well.
>
> "EnableRobotCloakActions = 1; "  caused our wiki to just show an HTTP
> error 500.
>
> This worked:
>
> wget      --recursive      --level 5      --no-clobber
> --page-requisites  -e robots=off --adjust-extension \
>      --span-hosts      --convert-links
> --restrict-file-names=ascii  --remote-encoding=ISO-8859  \
>  --domains    ourwiki.portal.se   --no-parent \
>   --no-host-directories     --reject-regex '(.*)\?(.*)'
> ourwiki.portal.se/FooWiki/SomeGroup
>
> The important bit is --reject-regex '(.*)\?(.*)'   which stops
> downloading of action= links.